Chapter 1. Writing for Web Users Implies Writing for Search Engines
Writing for the Web is fundamentally different than writing for other media, such as print. Sure, some good writing habits for print also apply to the Web. You should use an engaging tone and fresh word choices. You should organize your information clearly. And, most importantly, you should understand your audiences and write in ways that make it easy for them to understand your content. However, analysis shows that readers approach Web content far differently than print content. This book seeks to use this insight to provide a practical guide for Web writers and content strategists.
This book is about understanding the content needs of Web users to do a better job of presenting relevant content to them. It assumes a good working knowledge of how to write for print and therefore will not delve into the mechanics of quality writing. But it will focus on the distinction between how print and Web media differ, which requires some explanation of how the print medium works. From this foundation, we can understand how the Web as a medium differs from print. We can then develop practical guidance on how to do a better job of engaging Web readers.
You might be skeptical about this. Whether for Web or print, text is text, right? In this book, it is our job to counter this skepticism. In subsequent chapters we will cite numerous case studies and deep research into user behavior that clearly demonstrate how Web readers behave and why they do. For the time being, however, we ask that you suspend your skepticism so that we can introduce the content of this book. What follows is a brief sketch of the chapters in this book, which we hope will convince you to read on. We promise that by understanding what is covered in these chapters, you can truly master a field that is crying out for competent practitioners: Web content writing.
How the Web Medium Has Evolved from Its Print Origins
The basic difference between print and Web media is in the reader/writer relationship. In print contexts, you typically invite an audience to journey with you through your prescribed content path. The best print writers encourage their readers to surrender control and let the writer lead them by the hand through the material. Often, print readers will readily concede this control, trusting that the writer knows how best to organize and present information.
On the Web, readers (if we may call them readers for the present) will not cede control over the information path. They navigate through paths of their choosing, cutting corners and trying to get to the most relevant content as quickly as possible. On the Web, it is the writer’s job to provide a multitude of clearly marked paths, letting readers find the relevant nuggets of information that they seek. How to write to let readers sift through your content and find those nuggets is a considerable challenge that deserves a book of its own.
A particularly salient example of how Web writing differs from print is the way Web readers use search engines. Web users are impatient with content providers, because they can be. If they can’t find the information they’re looking for by navigating to it, they will use search engines. This impatience with information retrieval shows up in their reading habits. As a study of Web users by Weinreich et al. (February 2008) has demonstrated, Web “readers” do much more skimming and scanning than print readers. The study shows that on average, people spend 82% less time actually reading Web pages than they do when they read print pages, assuming average print reading speeds of 250 words per minute.
As Jakob Nielsen (June 2008) shows, Web users usually don’t read pages in the conventional way, line by line, serially. They scan for keywords in the heading and short descriptions and only read after deciding that some content is relevant. With this in mind, Google has designed its search crawler to mimic how Web users behave. The crawler scans pages for keywords and captures the pages with the strongest placement of those keywords to include in its index. When a user enters a keyword phrase into Google’s search field, Google returns the results that its algorithm deems relevant to those search terms. The design of its crawler is one reason that Google has become the search engine market leader in the United States and elsewhere. It tends to return highly relevant results for users, and it displays those results in ways that users can easily digest, given their extreme impatience. Because Google and other search engines strongly cater to Web user behavior, learning to write for the Google algorithm is an essential aspect of writing for Web readers.
Though our book relies on much of the information provided in Mike Moran’s and Bill Hunt’s excellent book Search Engine Marketing, Inc., Second Edition (2009), this is one point where our approach diverges from theirs. Hunt and Moran claim (2005, 309) that “The best philosophy for writing for search is: Write for people first, not for search engines.” Our claim is that writing for search engines approximates writing for people. Also, Web writers often lack audience knowledge—readers can come from anywhere using search. Because Web writers often lack audience knowledge, writing for search engines is often the best way to understand how to write for people. So we take a “write-for-search-engines-first” approach.
Writing Relevant Content for the Web Audience
How do you analyze your audience for print publications? Suppose you write for academic periodicals. If so, you have a good sense of the history of debate within each one. And you know that readers of a given periodical are professors or graduate students in the field. Perhaps you have a demographic survey of its subscribers. From this, you form a mental representation of a typical reader (maybe even someone you know, like your advisor), and as you write or review your own work, you imagine that person reading it. In the print world, this is the closest you will ever get to addressing audience members based on known facts about them.
In most print contexts, you know significantly less about your audience than you do in academic periodical contexts. Magazine writers might know rough demographics about subscribers, but they never know who might pick up a given publication at newsstands. Book writers know even less about their audiences. You might write a book for a particular audience, in the sense that you define its topic and purpose so that audience members who are interested in those things will buy it or check it out. But you don’t always know exactly how to address them, either as individuals or as a group. Print audiences are typically much more diverse than subscribers to an academic journal. It is simply not possible to address all possible readers with one print product. You just don’t know them well enough to do this.
For this reason, many print writers invoke their audiences by using storytelling and other compelling techniques to draw them into the book’s world. When readers start down the path of a particular story, they leave all expectations of being addressed for who they are at the trail head and follow the writer into unexplored territory. The more richly the writer creates that territory, the more readers will feel compelled to take the journey.
On the Web, you know a little more about your audiences, but your knowledge is fairly generic: You might know their service providers and perhaps which search engines they came from, and which paths they take through your site. But on an individual level, you don’t know much at all. (Unless they sign in to your site. But let’s leave those cases behind, since very few users take the time to do so. And even if they do sign in, what you know about them can’t help you tailor messages for them.) Because users take unpredictable paths through your information, you can’t connect with your audience as you do in print, such as by addressing a tightly defined audience or by appealing to a diverse audience. You have to find some way to connect with them in order to deliver content that they will find relevant. There are no perfect solutions to this problem, but we have developed some strategies and tactics to help you better connect with anonymous audiences on the Web. All of them center on search.
When you write content explicitly for search engines rather than for your users, in a sense, you invoke the search engine users with an effective mix of keywords and links that draws them to your pages. The challenge is to craft your pages in ways that attract specific users from search engines, especially Google. In so doing, you can present relevant information to your audience. This book is about attracting your audience with keywords and links and thereby providing relevant information to them. As mentioned earlier, because Google and other search engines cater to the way users scan and retrieve content on the Web, writing for Google is also an effective way to write relevant content for your audience.
Discovering and Using Popular Keywords
The first thing to do if you want to optimize your pages for search is to find out what keywords related to your theme or topic are most often searched for. These keywords become your site’s nomenclature. If you use these words in prominent ways on pages in your site, you will have a better chance to get traffic.
But traffic volume is not the end game. The end game is targeted traffic. You want to engage with your visitors. You want your target audience to come to your site and find that your content is relevant to them. Visitors who find your content irrelevant typically click the back button or “bounce” off your site without clicking any of its links. If you try to get high traffic without taking care to also target your audience in the process, you will get a lot of traffic; but most of it will bounce. What you want, instead, is low bounce rates with relatively high traffic. How can you achieve this? It’s not as easy as it might seem. Popular keywords that are used by many people in a variety of contexts will yield mixed results, if all you do is optimize your pages for those keywords. You will get high traffic volume, but also high bounce rates. The first step is to develop a set of related keywords—or a keyword cloud—which your target audience uses frequently. Then you need to develop pages that use the words in this cloud.
So how do you develop these keyword clouds? One way is to use keyword research tools to find related keywords. These tools can help you identify not only the most often searched-on words, but also related words and how often they are searched on. Once you get a sense for the number of users who search on a keyword—or its search demand—you can use the most relevant, high-demand words as the building blocks for your content.
When we use the term keyword, we do not merely refer just to single words. Most keywords that users enter into search engines consist of phrases. A keyword cloud typically contains not only related single words, but also related phrases. Many users search on so-called long-tail keywords to zero in on the exact content they are looking for. These are not just longer strings of words and phrases, though they typically are longer than high-demand single words. The phrase long-tail refers to the demographics of users who search for very particular content, rather than searching on more generic topics. Users who enter long-tail queries tend to be more search savvy. If your content ranks well for these, you will attract a highly targeted audience. But no single long tail will garner much traffic. The number of long-tail keywords in your content will need to be enough to drive targeted traffic to your site. For these keywords, you have to understand the language of your target audience at the sentence level, rather than at the phrase level. One way to develop this understanding is to research your target audience’s social media hangouts, such as blogs, communities, and forums. Because users tend to use the same sentences in their long-tail keyword searches that they use in social media contexts, knowing the writing habits of your target audience will help you know the best long-tail keywords to use.
There are a variety of tools you can use to better understand the writing habits of your target audience. Very effective ways to do this include Google Alerts, Yahoo Alerts, and a method that uses Yahoo Pipes to track mentions of your company via RSS feeds—subscription feeds that automatically update when the source content is changed. You can use these methods to guarantee that whenever someone mentions a particular phrase (such as your company name) in a blog post, you get an RSS notification and can look and see what that person has said. This not only helps you get a sense of how users feel about your offerings, but also about what kind of language that blog’s readers use for them. Later, we will show how to use free tools like Yahoo Pipes to monitor social media for common keyword-related activity.
Engaging with Web Visitors through More Targeted Search Referrals
Until now, we have focused on using keywords to attract a target audience to your content. But keywords are not the only parts of Web content that determine whether your content is relevant. A user can find the content on a page (with the same keywords) relevant one week and irrelevant the next. You might ask: How can that happen? If keyword usage determines relevance, how can attracting users to your pages though keywords drive users to information that they find irrelevant? Well, language is a lot more complex than creating a simple matching algorithm between keywords and users. If it were that simple, Web writing and editing would be a matter for technology and would not require human decisions. Fortunately, making good content decisions based on a variety of variables, including keyword usage, requires humans.
Suppose that a visitor to one of your pages has viewed all the information on it, but there have been no new updates since then. The content might still be relevant to the visitor’s interests, but no longer relevant enough to reread. In a marketing context, users might come to your site one week to see what you are offering, and the next week to see how those offerings fit their needs. Once they are aware of your offerings, if you simply drive them back to the same page through search, you’re creating an irrelevant experience—one that could end in a bounce and a bad user experience.
There are many more variables that affect relevance than we have space to list. Those are for linguists to determine, rather than writing instructors. But we can point you to one important variable beyond keywords that affects whether visitors will find the content on your pages relevant: purpose. If you tune your pages to the activities your users hope to accomplish when they arrive on your pages, you have a much better chance of getting them to engage with your content. Engagement is one way we measure relevance: If people click a link on your page, we can at least say that link is relevant to them. Sometimes engagement is merely a question of reading the content. Sometimes it’s clicking a link. Sometimes it’s getting users to comment or fill out a form. There are countless calls to action that a Web site can have—another key difference between print and Web. In print, you simply want to get your reader to read and comprehend your information. Perhaps you want the reader to be entertained or merely informed. However, you never want the reader to interact with a book—to write in the book in hopes that you will write back.
But on the Web, engagement or interaction is typically exactly what we want users to do. If all we want them to do is read and comprehend, we can provide PDFs for printing and reading offline. But good Web content is interactive. It compels the visitor to take action. So, in addition to tuning your page’s content for keywords, you also must tune it for the interactive purpose of that page. The real trick is mapping the keyword phrase to the page’s purpose. This isn’t as hard as it might seem. It is merely a question of adding a verb to your keyword phrase that describes what you want users to do when they get to a particular page. For example, one page can have keyword phrases with learn in them, another page might have keyword phrases with shop in them.
The question is: How do we help users land on one of our pages that is relevant to a keyword and related to their purpose in seeking the information? We will attempt to give some answers to this question in this book. But a word of caution: There are very few general answers. Different environments demand different answers to complex search questions. You will never achieve perfect engagement levels; because of the complexity of Web interactions, some visitors will bounce. But we can give you a framework for answering the question, and improving your engagement rates with writing that is more focused on the purpose or user goal of the page.
Developing an Optimized Site Architecture
We think of print publications as mostly self-contained units. Sure, we refer to related works in the bibliography or source list of a publication. But we expect print publications to be consumed whole. This is another key difference between print and the Web. Web users do a lot more skimming and scanning than print users. Only after they determine that the information matches what they’re looking for do they bother to read. This is a central insight of this book, and it affects every aspect of Web publishing, including design and architecture.
For our purposes, Web architecture is the practice of designing information experiences that help users find the information they’re looking for. In a sense, writing for search engines is part of this practice. If you write in a way that helps users find information more easily through search engines than if they navigated to it from a home page, you are approaching an optimal user experience. But search is not enough, either; you also need to design for navigation. The goal is to create engagement, and in some cases, conversions. It’s not enough to get a visitor to click a link; that click should land the user on a relevant page. Search can draw users to lower-level pages and encourage them to navigate up; or it can draw users to higher-level pages and enable them to navigate down. And horizontal navigation is also part of a good architectural plan. In any event, the content experience doesn’t end with getting the user to come and click on something; you must get the user to engage with it.
Many architectural discussions at IBM and other companies focus on designing a hierarchy of pages that enables users to easily move from the top level to the specific information they’re looking for. This is a good approach if you are designing a user experience for navigation. But it leaves out key considerations that can help users find information from search. Some search engines rely on metadata—extra-linguistic information hidden from view in the code of pages—to help determine the relevance of those pages. Though Google does not use metadata as part of its ranking, it does analyze how pages are interlinked with the rest of the Web to help determine relevance. Thus, architecting a set of pages for search engines necessarily includes paying attention to metadata and linking.
In this book, we focus on Google as the most popular search engine in the United States and elsewhere. From an architectural perspective, this means a thorough discussion of linking. This is the subject of an entire chapter later in the book.
Before we do that, we will first discuss how the standard practice of designing a hierarchy of related pages around a central topic or theme relates to keyword usage for search engines. To reach a more general audience, most architects design information around topics starting at a high level. This architecture should help users drill down into the information in the hierarchy, according to their individual needs. To enable this experience, you should choose different related keywords for each page in the topic, following the hierarchy. For example, on the IBM Web site, we might start with a high-level page on IBM Servers and then develop pages related to a specific product line, like the BladeCenter and other IBM offerings within this topic, such as BladeCenter hardware (Figure 1.1).
Figure 1.1. A hierarchy of pages in ibm.com.
You might think this is a fairly straightforward process. The architect designs the hierarchy of pages, and the writer picks keywords to fit into it. We suggest that this process rarely works the way it is drawn up on paper. At IBM, we struggle with pages in a hierarchy that do not produce search referrals because users simply do not search on the keywords we chose for those pages. Imagine a hierarchy of pages in which the third level down the tree is the highest ranking page because it uses a popular keyword, yet the top-level page in this hierarchy gets little or no search engine traffic because few users ever search on its keyword.
We suggest (and will demonstrate later in the book) that the best practice is to optimize the top pages of the hierarchy with the most competitive keywords—the ones that draw the broadest audience. Pages lower in the hierarchy need to be optimized for narrower, more targeted audiences, who typically use long-tail keywords. You do this by choosing related keywords that are more likely to appeal to specific segments of the broader audience. The point of our chapter on the relationship between architecture and writing is that architecture reflects writing practices: When we create search-first architectures, we do a better job of creating an information experience for users for the whole hierarchy, not just for isolated pages within it.
The goal of information architecture is to serve users with relevant information. But how do you know what information is relevant to them? We suggest that the search-first architecture does a good job of creating relevant information for a large set of users, with less guesswork and less trial-and-error than common architecture processes. Keyword data is the best information available from which to design your information. It not only helps you isolate specific keywords that will draw higher volumes of users to specific pages, but it helps you understand how keywords and phrases are related to one another. If you optimize a hierarchy of pages for a set of related keywords, you not only direct users to specific pages they might be interested in; you can also get them to navigate to other pages in the hierarchy from their initial search referral. If they find the top-level pages relevant, chances are they will find pages targeted toward more narrow audiences even more relevant.
In large organizations with a complex matrix of Web sites, content creators can unwittingly compete for the same keywords, thus harming visitors’ overall experience. For this reason, large organizations need to manage keywords across their whole site, not just within specific areas.
Suppose you own a part of a company’s Web content—perhaps the marketing pages related to a portion of your company’s product portfolio. You do all the keyword research related to that portfolio, develop a site architecture that maps pages to popular keywords and desired visitor interaction, and write optimized content for each page in the architecture. In short, you do everything right to attract the most targeted audience possible to your pages. However, suppose that a colleague owns the Web content for a related set of offerings in your company’s portfolio. She does everything you do and optimizes her pages for the same keywords, all the way down to the long-tail ones. You are now unwittingly competing with your colleague for the same targeted audience.
As unlikely as this scenario is, competition is actually quite common in a company such as IBM that has a large and diverse portfolio of offerings, and many Web pages related to them. Even for themes such as Green IT, several efforts might spring up at the same time and could unwittingly compete with one another for the same users, unless these efforts are coordinated. For this reason, we recommend corporate-wide keyword management systems, which enable content owners to reserve specific keywords for specific pages. These systems can spread keyword usage across an enterprise in way that is similar to, but more pervasive than, what you do when you develop a keyword-based architecture around your theme or topic. With such a tool in place, you can optimize your enterprise for popular keywords and attract targeted audiences to the most important pages for your business.
Gaining Credibility through PageRank
As important as keyword usage is for your search efforts, it is less than half of the Google algorithm. The Google algorithm takes two primary things into account: relevance, which is a function of how keywords are used on a page; and PageRank, which is a function of how your page is interwoven into the Web’s vast map of links. (There are other factors as well, such as the prominence of your site in the scheme of the Web, but those are beyond your influence as a Web writer. In this book, we will focus on keyword relevance and PageRank.) You can do everything right to improve the relevance of your pages by using keywords, yet still fail to get your page listed on the first page of Google results. If a particular keyword is highly competitive, meaning that lots of other sites are doing everything in their power to rank high for it, you will also need to improve your PageRank to get listed on that first page of Google, Bing, and other search engines.
For example, the keyword Service Oriented Architecture, or SOA, is highly competitive in Google—a lot of very smart companies spend a lot of money developing optimized pages around a cloud of keywords related to SOA. The only way to get on the first page of results in Google for SOA is to get credible pages around the Web to link to your SOA page. Google counts each link to your page as a vote of confidence for the content on your page. This is how Google overcomes the problem of relevance on the Web. Because the text alone cannot determine whether users will find the content relevant to them, and there are no better contextual cues to content than links, Google uses linking as its main contextual cue. All evidence indicates that Bing also uses links, and, if anything, gives them even greater weight than does Google.
As complicated as keyword usage can get, linking is much more complex. Search engines don’t count every link to your pages as equally valuable. Some links get more link juice, or value, than others, depending on how valuable and relevant Google deems the site that links to your page.
We will explain some of this complexity later in the book. For now, suffice it to say that you need to develop a plan to promote your pages to high-value sites. Many sites have value simply because they serve as link aggregators related to a given topic. One example is Wikipedia: It contains relatively little original content, but it is an excellent place to begin researching a topic, because every source is a link to further research. (If you attempt to publish original content on Wikipedia in the sense that it lacks original sources, the content gets flagged as in need of support.) However, for technical reasons to be explained later, Wikipedia is not a great site from which to get link juice. It merely illustrates the point that sites like it, which link to deeper dives into research topics, are highly valued by users. And because users value them highly, search engines do as well. For this reason, these hubs of authority typically find themselves listed near the top of search results.
Our approach is to determine the best path to becoming a hub of authority on your topic. This requires a lot of deep thinking about the state of the art of your topic and a lot of research into the existing hubs of authority on it. How does your content fit into the matrix of authority on your topic? What contribution does your content make to the interlinked conversation related to your topic? How do you engage with other subject matter experts in your field to promote this contribution to the discussion? We will explore these and other questions to help you improve your PageRank in Google. And who knows, if you develop a good relationship with the recognized authorities in your field and your content makes a unique contribution to the field, you too can become a hub of authority and thus rank highly in search engines.
There are no shortcuts to becoming a hub of authority. To do so, you must gain the trust of other authorities for your topic. That takes time. Still, there are some tactics that can help you promote your content to authority sites. We will cover them in Chapter 8.
Capturing Social- and Rich-Media Opportunities
Print habits die hard. Chief among them is the habit of wanting to control the conversation. When you write for print, it’s your tale and you control the telling—you control how the reader consumes the information. But when you write for the Web, the reader controls the pace and flow of information. Trying to control it is a surefire way to get users to bounce off your pages. The best you can do is to give users options they will want to choose. Part of gaining their trust to choose your content options—to click your links—is making it clear that it’s their choice. Part of this is demonstrating that your content is not isolated, that it doesn’t claim to have all the answers, and that it is but a small part of a bigger conversation.
The Web is an evolving medium, and users’ expectations evolve with it. At this time, the fastest growing practice on the Web is sometimes called social media. Loosely defined, it is a set of practices that engage Web users to participate in the conversation rather than merely consume static information. These practices include blogs, wikis, forums, persona sites such as Facebook, microblogging sites such as Twitter, and the like. Social media is rapidly evolving to include graphics and video sharing and sophisticated syndication.
Nowhere is the Web more distinct from print publishing than in social media contexts. Users control the flow of information as they navigate through static Web pages; but they actually contribute to information in social media contexts. Here, their control of the information is complete. In extreme cases, users who make frequent comments on a blog are as important to the blog’s success as its author. And a blog is not much more than an online opinion column if no one comments on its entries.
As the Web becomes more of a collaborative medium—more of a space for symposiums rather than lectures—users’ expectations for the whole Web change. Even publishers of static Web content need to adjust their practices to engage users in conversations. Whether your site explicitly enables users to comment on its content or not, users expect it to. Social media has accelerated the need to give users more control over their information paths, even if your content is not intended to be shared and commented on.
For example, at one time IBM had a number of Web producers who preferred to get permission to republish content from other sources on their sites, rather than to link directly to the sources of that information. Fortunately, across the industry, this practice has long ago become a distant memory of how things used to be done in Web publishing. That memory is all that remains of writing habits that stem from print: “Let’s keep users on our site. Let’s not let them leave.” This attitude is self defeating on the Web. You lose users’ trust when you try to control their experiences; and you lose potential PageRank, as well.
Because page rankings are a large part of Google’s PageRank (and of Bing’s similar algorithm) and are determined by how a community “votes” on the credibility of content, search is a social medium. Not surprisingly, improving your standing in the community by engaging users in social media contexts is a very effective way to gain credibility and PageRank. If you encourage users to pass your links, podcasts, and videos around to their friends, the PageRank for the pages on which those assets are accessed will grow virally. Since users expect this pass-along content to be ever more visual, search effectiveness is not just about text anymore. It is crucially about rich media as well.
The irony is that the more control you give users, the more they will want to visit your site. The more open you are in your reciprocal linking relationships to and from rich media on other sites, the more users will treat your site as a hub of authority and will keep coming back to hang out and explore new paths through your content. Later on, we will demonstrate in more detail how to take advantage of the social attitudes and habits of Web users.
This letting go of control of your information is a necessary, though sometimes painful, cultural shift in Web content practices. Because this shift is so necessary and fundamental to successful Web content efforts, we believe it deserves a chapter all its own.
Measuring Web Effectiveness
How do you know how well you are doing on Google? That is a rather easy question: You simply search on your keywords and see where your page ends up in the results. How do you know the volume of visitors you get from Google? This question is a little harder: You use a Web analytics tool such as Unica Netinsight to find out where your visitors are coming from and then filter the results to show only those who come from Google. You can also run reports in modern Web analytics tools that show what keywords brought users to your pages, and in what volume.
The more detailed you want to get in understanding the effectiveness of your search efforts, the harder the questions become. You can get a lot of information related to data such as Google ranking, Google referrals, raw traffic, no-results keywords, click throughs, and customer feedback specific to search, but correlating that data is quite difficult. The goal of this kind of research is to measure engagement: How many users are doing the things you want them to do on a page, such as downloading a white paper, filling a survey, taking a poll, or making a comment? We will discuss a best practice in gathering and correlating the various search effectiveness later in the book.
For the purpose of this discussion, suffice it here to say that how you define and measure search effectiveness is a fundamental facet of using search to target audiences with content. Search efforts rarely work exactly as they were drawn up on paper. More often than not, you will find aspects of your search efforts that underperform expectations. In these cases, you will have to go back to your pages and fine-tune your content to better target your desired audience. If you find a high bounce rate, for example, you might need to change the language on the page to develop better engagement with your target audience. The practices outlined above and detailed later in this book can help you do a better job of search optimization when you first develop and publish content. But language on the Web is a complex medium, and you can expect to have to do a certain amount of reworking to achieve better engagement. For this reason, measurement is a critical phase in search optimization work.
Also, to compete for users, search engines are evolving rapidly. If an engine changes after you publish information, it might make sense to go back and tweak your pages in ways that can help you rank better for it. To do this, you need analytics tools and best practices. We will describe how to use available Web analytics tools to measure search effectiveness and develop action items to improve existing content in the process.